Search papers, labs, and topics across Lattice.
1
0
3
Multimodal models stumble badly on low-resource Southeast Asian languages, as revealed by the new SEA-Vision benchmark for document and scene text understanding.
